1. Field of the Invention
The present invention concerns in general terms a method of allocating transmission resources for a mobile telecommunication system, more particularly for a system of the TD/CDMA type, for example a telecommunication system of the UTRA-TDD type.
2. Discussion of the Background
In the UMTS (Universal Mobile Telecommunication System), the access network, known as UTRAN (standing for UMTS Terrestrial Radio Access Network), provides the connection between the mobile terminal of the user (MS, standing for Mobile Station) and the interconnection network (CN, standing for Core Network). It includes access network controllers (RNC, standing for Radio Network Controller) and base stations (B nodes), each RNC controller being able to control a plurality of base stations. The downlinks between the base station and the mobile terminals which come under it are separated from the uplinks either by the use of different frequencies, a mode known as FDD (standing for Frequency Division Duplex) or by different transmission time ranges, a mode known as TDD (standing for Time Division Duplex).
FIG. 1 depicts schematically an example of a transmission frame in a UTRA-TDD telecommunication system (UMTS Terrestrial Radio Access Network in TDD mode). The transmission frame has a total duration of 10 ms and is divided into 15 transmission time slots denoted TS0 to TS14. The transmission frame is also divided into at least two distinct time ranges (duplex partitioning) and allocated respectively to the uplinks (represented by up arrows) and the downlinks (represented by down arrows). The separation between consecutive time ranges is referred to as the switching point. A frame in general has several switching points. In order to reduce interference, adjacent base stations are synchronous with each other and have identical switching points. The separation of the different uplink (or downlink) communications within the same cell is provided by spreading sequences (known as signatures) using orthogonal codes. The communications, up or down, within a cell are isolated from the communications of an adjacent cell by means of codes known as scrambling codes. In order to reduce intercellular interference, the jamming codes relating to adjacent cells are chosen so as to have a low level of intercorrelation. However, because in particular of the scattering of the delays caused by the multi-path propagation, this low intercorrelation of the scrambling code is not in practice satisfied on reception. As a result the data transmitted at a given time within a cell may be interfered with by data transmitted at the same time within an adjacent cell. This intercellular disturbance or interference is variable since it depends on many factors, amongst others the spreading codes, the scrambling codes, the transmission powers and the characteristics of the different transmission channels of the different users. Hereinafter, the combination of a spreading code and a scrambling code will simply be referred to as a code.
In general terms, a telecommunication system of the TD/CDMA (Time Duplex/Code Division Multiple Access) type will be considered, that is to say a system with multiple access division which is both of the time and code type. Each communication is allocated a time resource (a time slot within a frame) and a multiple access resource (a code) at this time slot. The codes are organised in groups, distinct groups of codes being allocated to adjacent base stations and the codes within the same group being orthogonal to each other. In other words, the partitioning into groups of codes serves to separate the communications of two adjacent cells and the partitioning of the codes within the same group serves to separate the communications within a cell.
In order to combat intercellular interference, a dynamic allocation of the transmission resources available at the base station has been proposed, referred to as FDCA (Fast Dynamic Channel Allocation).
Various dynamic allocation strategies are known in the state of the art.
It will be assumed hereinafter, in order to simplify the disclosure, that only a set of uplink communications or a set of downlink communications is involved. It is clear, however, that the dynamic allocation method must operate on both independently.
A first dynamic allocation technique uses a criterion of the lowest interference level. According to this technique, a base station regularly transmits a measurement request to the mobile terminals which it serves. On reception of such a request, each mobile terminal makes an interference level measurement and transmits the result of it to the base station. The base station also measures the interference level within each time slot. When a new communication is to be established, the base station chooses, from these measurements, the time slot having the lowest level of interference. The communication is then established on the chosen time slot and on a free code within this slot. The consequence of choosing the transmission slot having the lowest level of interference is that the transmission power necessary for the new communication will not be too high. Moreover, the additional interference produced by the latter will also be low. In this way there is obtained a mean level of interference which is lower than that which would result from a random time slot allocation. FIG. 3 gives an example of allocation according to this method. The left-hand diagram depicts the interference level measured for the different time slots. The time slot of index i having the lowest interference level is adopted for the allocation of resources to a new communication. The right-hand diagram expresses the number of codes allocated according to the time slot. The hatched block illustrates the resources allocated to the new communication.
A second known dynamic allocation method uses a criterion of condensation of the time resources. It consists of allocating, to a new communication, the time slot (not saturated in terms of codes) having the largest number of codes allocated. In this way there is obtained a maximum concentration of communications on a minimum number of time slots, which releases the largest number of time slots available. The codes within the same slot being orthogonal, the signals of the different communications do not interfere with each other, in so far however as the transmission channels do not interfere with this orthogonality. The cells adjacent to the cell in question can thus be allocated time slots which are not occupied with codes not orthogonal to those used in the said cell. This technique a priori reduces both the extra-cellular interference, since adjacent cells use distinct time slots, and intra-cellular interference, since the communications within the same cell use orthogonal codes to the maximum possible extent. FIG. 4 gives an example of allocation according to this method. The left-hand and right-hand diagrams show the number of codes allocated according to the time slot, before and after the allocation of resources to a new communication. Cmax indicates the maximum number of codes available per slot. It can be seen that the slot of index i is allocated since it is the one which has the largest number of codes already allocated whilst still offering an available code.
The two allocation techniques described above both have drawbacks.
This is because, according to the first technique, two mobiles which are close but situated in two adjacent cells may be allocated the same time slot since the latter is, for each one, the one which has the lowest interference level. However, the codes which will be allocated will not be orthogonal to each other since they belong to distinct cells and therefore to distinct groups of codes. Paradoxically, these two mobiles which will use the transmission slot which is a priori the least interfered with will be in a situation in which they mutually interfere with each other to a significant extent.
As for the second technique, it in fact requires cellular planning of the time resources. For two adjacent cells, the condensation of the time slots must take place in distinct slots. This is because condensation in a common slot would result in catastrophic interference between the communications of the two cells using this slot, all the more so if it were heavily occupied. However, it is known that such cellular planning is not only complex to implement but in particular results in sub-optimum use of the resources at the level of a set of adjacent cells: a cell with low traffic adjacent to a cell with high traffic could in fact unnecessarily immobilise resources needed by the latter.